Photo: respresSuppose you were polishing the hood of your car and accidentally summoned up a transportation genie from the depths of your transmission. What would your three wishes be? (Guys: assume this is a genie who looks more like Shaq than Barbara Eden.)
How about:
- Cure congestion. Allow cars to race at 65 mph down the Jersey Turnpike or the 405 at the height of rush hour, with no new blacktop or tolling necessary.
- Increase safety. Even with those higher speeds, cut fatal accidents way down. And achieve this even with lots more dangerous behavior, like in-vehicle texting, drunk driving, Big Mac eating, YouTube watching, and the always-addictive computer solitaire.
- Improve mobility for the transportation-deprived members of our society: the young, the elderly, and the disabled.
Are these developments in the realm of fantasy? Happily, no. In fact, a technological suite is progressing which will allow us to enjoy all three simultaneously, meaning you can save your wishes for Middle East peace, tomorrow’s lottery numbers, or cursing the Miami Heat.
The miracle innovation is self-driving, robotic cars. Urban planner (or, as he fancies himself, “antiplanner”) Randal O’Toole has a nice chapter on this looming development in his recent, informative and provocative book Gridlock. (More on the interesting and controversial O’Toole coming up.)
Driverless cars may seem the stuff of science fiction. Futurist Norman Bel Geddes predicted we’d have them in a couple of decades – in 1940. They’re not here yet, but, hey, in 1940 nobody believed wristwatches that let us talk to each other on TV would make it out of the Dick Tracy comic strip, and today they are basically reality. So are computer systems which know where we are and can direct us to where we are going anywhere in the country, which certainly would have seemed fantastic just a few decades ago.
At the moment, robot cars are not the subject of much public scrutiny or debate. However, that should change given that important technological innovations are not just feasible but are currently being deployed on the roads.
As you may know, some new Lexuses and Toyotas parallel park themselves, a great boon to all except residents of New York and San Francisco, where there are never any curb spaces to begin with. The problem of cars navigating in more complex urban situations is orders of magnitude more difficult, but a solution is on the horizon – in 2007, in the DARPA Challenge several robot cars successfully negotiated a demanding 60 mile urban course.
Many of you might be disturbed by the safety implications. It is probably safe to assume that robot cars will not rise up against their masters and enslave us all, but any PC user undoubtedly has concerns about computer crashes leading to real ones.
This certainly must be taken seriously, but, after all, new cars are loaded with microprocessors that control things like your fuel injection and transmission, and which function so flawlessly that you don’t even know they’re there. And the next time you’re on a plane, consider how you’re placing your life in the hands of a computer (the autopilot, which even basically lands the plane in bad weather) without giving it a second thought.
Moreover, we have to consider the shortcomings of the auto control devices computers would replace – us. Human beings are admirably adapted for replicating our genes: it’s doubtful any computer could look as good as we do on the dance floor. But we are poorly adapted for driving cars, which move at speeds far above what humans ever achieved prior to 1800.
Thus more than half of the fatal accidents in America — which claim 40,000 lives a year – are due to driver error. When perfected, computers would be so far superior to us – in terms of reaction time, consistency, perception, focus, etc. – that the safety benefits of self-driving cars would greatly outweigh any new safety concerns.
At this point, computers are enhancing safety, not undermining it. A growing number of high-end new cars feature “adaptive cruise control,” which utilizes radar (more technologically advanced, but costing between $1000 and $3000) or lasers (less effective, but much cheaper at $400-$600) to judge the distance to the car in front of you, automatically cutting speed when you are getting too close for comfort. Volkswagen, Raytheon and other companies are also developing anti-collision radar systems.
Another innovation that has become reality (thanks to automakers like Nissan, Toyota and Honda) is the “lane keep assist system.” Here cars sense the position of the lane markers and resist efforts by the car to drift unless a lane change is signaled.
Because drivers will pay for it, these innovations are focused on safety. But equally importantly, driverless car technology holds the seeds of congestion relief.
Humans can judge that the car in front of them is slowing down (its brake lights flash and the car is growing), but we are poor judges of how quickly it is slowing down. Thus, to err on the side of safety, we have to overapply the brake. This creates a chain reaction when the same behavior gets replicated by driver after driver down the line. Traffic thus slows far more than necessary.
Slow reaction times and poor perception also mean that we have to leave an unnecessarily large amount of distance between us and the car in front of us.
With a system in which cars cooperate and inform each other (and perhaps the road) about their intentions instantaneously, they can start and stop relatively simultaneously. Moreover, cars can be packed much more tightly together. The outcome: a freeway lane can hold between three and four times more cars than are currently accommodated. And those cars can move at high speeds. Xi Jou of the University of Minnesota has estimated that congestion relief will start to bite when as few as 20 percent of the cars on the road have adaptive cruise control.
Finally, driverless cars hold the potential for revolutionizing the lives of those who are currently deprived by our auto-centric transportation system. A 14-year-old would be freed to go to school or the mall without an onerous and time-wasting journey by an adult, and a 90-year-old would no longer have to face an agonizing decision about giving up her license.
Despite the fact that we are making real progress – in 2008, General Motors announced it was planning a driverless car that could be on the road by 2018 – the difficulty of an automated traffic system is, of course, in the implementation. Many of the regulatory, logistical, financial and technical problems involved are quite vexing. They will require political will. More coming up on what we can be doing to put this issue on society’s radar screen.
and 
. You can also watch it on 
The robotic tractor trailer truck will happen way earlier than the robot car since transportation companies have so many $100s of millions to save if they could layoff truck drivers…
How about just much smaller cars? Tax the externality, and force people to drive much, much smaller cars? Want to drive? Fine, you get a smart-car size (or smaller) footprint on the road. Anything larger than that, and you’re taxed exponentially.
I think about this almost every day when I see thousands of my fellow Atlantans driving by themselves on 285/ The Perimeter with a car taking up space for 6 people.
In Tom Vanderbilt’s book, I believe he talked about if only 10 percent of cars had some speed/ distance regulating software, that it would be enough to make a difference.
But, in the end, the answer isn’t in how to make the automobile more efficient, or travel more coordinated. It is about how to reduce our overall “need” for such big vehicles in the first place.
“…And the next time you’re on a plane, consider how you’re placing your life in the hands of a computer (the autopilot, which even basically lands the plane in bad weather) without giving it a second thought…”
This is a wild exaggeration. Spoken by a true non-pilot who apparently knows nothing about how planes fly, or what autopilots (as well as the flesh and blood variety) actually do and don’t do.
Exaggerations about cockpit automation are so rife, and so entrenched, that they’ve become accepted as fact. In reality the number of commercial flights that are landed “automatically” by the autopilot system is well under one percent.
As an airline pilot I see only a small handful of automatic landings — “autolands,” in the vernacular — every year, not counting those practiced in the simulator. Practiced? Yes, because in most respects automatic landings are more complicated, and more work-intensive, than those performed by hand.
This is something I’ve written about extensively in my articles at Salon.com. Try here, for example…
http://www.salon.com/technology/ask_the_pilot/2010/06/24/planes_don_t_fly_themselves/index.html
Conventional wisdom holds firmly the idea that pilots are on hand merely as a backup, that we “monitor” rather than fly. The press and pundits repeat this nonsense constantly, and millions of people now believe it.
Patrick Smith
A motor vehicle that I don’t have to drive myself? I believe we have those, and we call them buses.
“…And the next time you’re on a plane, consider how you’re placing your life in the hands of a computer (the autopilot, which even basically lands the plane in bad weather) without giving it a second thought…”
Here is the longer version of my response…
“…And the next time you’re on a plane, consider how you’re placing your life in the hands of a computer (the autopilot, which even basically lands the plane in bad weather) without giving it a second thought…”
I fly for a living, and I write about air travel, and both pursuits I do a fair bit of mythbusting. Air travel has always been rich with conspiracy theories, urban legends, wives’ tales and other ridiculous notions. Nothing, however, gets me sputtering more than the myths and misunderstandings about cockpit automation — this pervasive idea that modern aircraft essentially fly themselves, with pilots on hand merely as a backup in case of trouble.
It’s the belief of too many people that a pilot’s job involves little more than watching the aircraft “fly itself.” In some not-too-distant future, the wisdom goes, pilots will be engineered out of the picture altogether.
This is so ridiculous that it’s hard for me to get my arms around, and it amazes me how often the pilotless planes conversation turns up — in magazines, on television, in the science section of the papers.
One thing you’ll notice is how the sources for these stories tend to be aerospace academics — professors, consultants, researchers — and not pilots. It’s not necessarily an academic’s job to be realistic or practical, and clearly these people are bright and knowledgeable. But they are often highly unfamiliar with the day-to-day operational aspects of flying planes, ignorant of the boundless operational contingencies of commercial flight — things that that no electric box can, now or in the foreseeable future, be wired to appraise.
Though pilots too are occasionally part of the problem. “Ah, this plane practically lands itself,” I heard a 777 captain once explain to a passenger. technophiles we enjoy boasting of the grand and special things our huge machines can do. And in our efforts to make things understandable for the layperson, we wind up painting a caricature of what flying is really like — at the same time undermining our value employees.
One evening I was sitting in economy class when our jet came in for unusually smooth landing. “Nice job, autopilot!” yelled some knucklehead behind me. Funny, I concede, but wrong. It was a fully manual touchdown, as the vast majority of touchdowns are.
Sure, most jetliners are certified for automatic landings — “autolands” in pilot-speak. But they are very rare, and the fine print of setting up and managing one of these landings is something I could talk about all day. If it were as easy as pressing a button, I wouldn’t need to rehearse them twice a year in the simulator or need to consistently review those tabbed, highlighted pages in my manuals. The technology is there if you need it — for that foggy arrival in Buenos Aires with the visibility sitting at zero — but it’s anything but simple. In most respects, an automatic landing is more challenging, more complicated, and more work-intensive than a manual one.
The analogy I’m fond of making is the one about aviation and medicine. Out in the field, automation helps a pilot in the same way that it helps a surgeon. It makes flying easier, but it does not make it easy. The technology is advanced and expensive and ultimately engineered to keep your customers safe and alive. But to understand how this equipment works, and to use it properly… well, you still need to be a doctor, or a pilot, first. You might here a surgeon make a comment about the “simplicity” of a certain procedure or operation, just as a pilot will sometimes speak of an airplane that is “simple to fly.” That in no way implies that the layperson could give it a go and be successful, and it does nothing to diminish the knowledge and experience required to perform at that level in the first place. Simple yes, but only in the context of a prerequisite level of expertise.
A flight is a very organic thing — complex, fluid, always changing — in which subjective decision-making is constant and critical. You’d be surprised how busy the most automated cockpit can become, and if I were to take even the most routine and trouble-free flight, from the preflight planning stage to block-in at the destination, and break it down event by event, explaining each of the hundreds of decisions and inputs made by the crew, big and small, from rote procedure to the unexpected judgment call, I would be typing for the next nine days.
Long and short: a jetliner can take itself from waypoint to waypoint along a preprogrammed route — a basic, skeletal outline of the flight. But the idea that a jet will “fly itself” to the destination without meaningful input from the crew is preposterous and downright offensive to anybody who flies for a living. That includes me, and sure that makes me an advocate. But it also gives me a very healthy sense of exactly what cockpit automation does and doesn’t do.
The technology is here. Car companies are incredibly slow to implement it because they fear the legal system. They fear that as they take more and more of the driving away from the driver, they become more and more responsible, ergo, liable, for what happens.
Yes, it might increase safety, but look at all the legal issues car companies have already faced over safety measures (especially things like airbags), and there comes a point where increased safety, even if paid for by the consumer, runs the risk of costing the car companies more than they’re willing to part with.
A fertile proving ground which will also saves us some money are the subway systems. Over rigidly define routes, minimal number of obstacles, well defined lightings, you have a much easier problem to solve. Once this is solved, one can redirect the energies of the subway drivers elsewhere. Similarly for the railroad systems. Once those are automated and people come to trust the technology, then there is a possibility for implementation on personal autos.
An interest legal issue suggests itself though: if a car on auto-pilot nevertheless hit a pedestrian, who would have legal liability? the human driver? the car manugacturer? the auto-pilot software firm? or the pedestrian?
The problems with robotic cars are more legal than technical. A robotic car could easily be ten times safer than a manually driven car, but the lawsuits for the remaining ten percent would drive any company into bankruptcy. Even when there was no proven fault the jury awards would be ruinous.
Lawyers almost destroyed general aviation in the US and even after Congress passed legislation to protect the industry, lawsuits are still is a millstone around the neck of the industry that hampers it’s growth.
That doesn’t mean that robotic cars won’t happen, but it probably happen in Asia and Europe long before we see them in the United States.
You might want to check out this blog on robocars. http://www.templetons.com/brad/robocars/